Francisella species (spp) are non-motile, pleomorphic, gram-negative, strictly aerobic, facultative intracellular coccobacilli. They are extremely infectious, as less than 10 bacteria are required for infection (Jones et al., 2005; Soto et al., 2009; Kamaishi et al., 2010). One member of the genus, Francisella noatunensis, 
has been reported worldwide as a cause of francisellosis in fish (Kamaishi et al., 2005; Mauel et al., 2005; Olsen et al., 2006; Mauel et al., 2007; Birkbeck et al., 2007; Jefferery et al., 2010). F. noatunensis is composed of two subspecies adapted to different host temperatures, one of which (F. noatunensis ssp. orientalis) causes disease in fish living in warmer waters (Kamaishi et al., 2005; Mauel et al., 2005; Mauel et al., 2007; Jeffery et al., 2010) while the second (F. noatunensis ssp. noatunensis) causes disease in fish living in colder waters (Nylund et al., 2006; Olsen et al., 2006; Birkbeck et al., 2007). Outbreaks of francisellosis in fish aquaculture can be devastating, causing large losses worldwide (Kamaishi et al., 2005; Mauel et al., 2005; Olsen et al., 2006; Mauel et al., 2007; Birkbeck et al., 2007; Jefferery et al., 2010), and represents the main challenge for aquaculture based on Atlantic cod Gadus morhua L. It is also sporadically problematic in aquaculture base on Tilapia, one of the largest produced fish worldwide.
Piscirickettsia salmonis is described as non-motile, not-encapsulated, pleomorphic coccoid, with a size ranging from 0.1-1.5 um (Mauel and Miller, 2002, Vet Microbiol, 87:279-289). Salmon Rickettsial Septicaemia (SRS), caused by P. salmonis, is a disease of salmonid fish with a huge impact on the salmonid fish farming particularly in Chile. Similarly to Francisella sp., also P. salmonis is intracellular in nature making vaccine development challenging. The mortality rate of affected fish varies, from more than 90% mortality in some Chilean outbreaks, to low levels of mortality in e.g. Norway. The reason for the observed differences in mortality is not known, and although environmental factors must be taken into account, strain difference is also likely. As annual losses due to SRS in Chile are estimated to be more than 200 million USD each year, the potential impact on the salmon aquaculture could be devastating. Despite the availability of several commercial vaccines against SRS with reported good efficacy in laboratory trials (Wilhelm et al., 2006, Vaccine, 24:5083-5091; list of available vaccines are reviewed in the Australian Aquatic Veterinary Emergency Plan, Disease Strategy Piscirickettsiosisl , 2013), SRS was reported as responsible for 60% of the mortality of salmon and 79% of the mortality rainbow trout in Chile in 2011 (Integrated Annual and Sustainability Report 2012, Cermaq: EWOS Innovation-SRS project in Chile). Clearly, there is a demand for a vaccine against SRS with improved efficacy. Up until recently, one of the main challenges within P. salmonis research and vaccine development has been the lack of growth of the pathogen in liquid culture media. Yañez et al., (2012), reported the AUSTRAL-SRS broth, a highly complex medium consisting of a marine-based broth supplemented with L-cysteine, that supported the growth of P. salmonis reaching an optical density of approx OD600 mn=1.8 after 6 days incubation. Improvements of growth rate and increased biomass was made by growth in basal medium 3 (BM3) reaching an OD600 nm of 1.7 after 37.5 hrs (Henriquez et al., 2013). BM3 consist of yeast extract (Merck) 2.0 g L21, peptone from meat (peptic digested, Merck) 2.0 g/L, MgSO4*7H2O 0.1 g/L, K2HPO4 6.3 g/L, NaCl 9.0 g/L, CaCl2*2H2O 0.08 g/L, FeSO4*7H2O 0.02 g/L and glutamic acid 2 g/L.
Aquaculture is able to prevent outbreaks of many bacterial infections that presented huge problems for the industry in its youth, by the use of vaccines composed of inactivated in vitro cultured whole-cell bacterial preparations (bacterins) supplemented with adjuvants (reviewed by Brudeseth et al., 2013). As a consequence of this, the use of antimicrobials in Norwegian aquaculture has declined enormously despite a large increase in the amount of fish produced (reviewed by Sommerset et al., 2005). No commercial vaccine for fish francisellosis is currently available (reviewed by Colquhoun & Duodu, 2011; reviewed by Brudeseth et al., 2013), as attempts at using whole-cell preparations of F. noatunensis ssp. noatunensis has yielded none or unsatisfactory levels of protection (reviewed by Colquhoun & Duodu, 2011). This is similar to the situation for tularemia in humans, where vaccination with killed bacteria induces an antibody response with only limited protective properties (reviewed by Cowley and Elkins, 2011). The reason for this is due to the fact that to develop proper protection against Francisella spp. there is a need to stimulate cell-mediated immunity (reviewed by Cowley and Elkins, 2011), which vaccines based on killed whole-cells or protein subunits generally are poor at (reviewed by Titball, 2008). Live attenuated vaccines (LAVs) are efficient at inducing cell-mediated immunity, though there are safety concerns such as reversion to virulence and safety for immune-compromised individuals for such vaccines (reviewed by Titball, 2008). Particularly in an aquaculture setting, spread of genetically modified organisms to the environment is another factor to take into account. A LAV designated Live Vaccine Strain (LVS) has successfully been utilized to protect high-risk groups against tularemia (reviewed by Conlan & Oyston, 2007), demonstrating that it is possible to generate successful LAVs against Francisella spp. Several targeted deletion strains have also been shown to be protective against tularemia, such as the F. tularensis ssp. tularensis Schu S4 Δftt_0918, Δftt_0918 ΔcapB and ΔclpB (reviewed by Conlan & Oyston, 2007, Conlan et al., 2010). Additionally, a LAV based on a ΔiglC mutant of F. noatunensis ssp. orientalis has recently been patented (U.S. Pat. No. 8,147,820 B2) for use against francisellosis in aquaculture, and has been shown to protect tilapia against experimental challenge (Soto et al., 2011). Previous ΔiglC mutants have been shown to induce protective immunity in mice for F. tularensis ssp. novicida but not against F. tularensis ssp. tularensis (Twine et al., 2005; Pammit et al., 2006). Protection obtained by vaccination with both F. tularensis ssp. novicida and F. noatunensis ssp. orientalis ΔiglC mutants in mice and tilapia respectively could partly be transferred by passive immunization of naïve animals (Pammit et al., 2006; Soto et al., 2011).
When constructing vaccines for immunization of Atlantic cod there are certain peculiarities of the cod immune system that should be addressed. Vaccination results in production of lower levels of specific antibodies and less variety in the utilization of immunoglobulin heavy chain types, but despite this Atlantic cod develop protective immunity after vaccination with most bacterial pathogens (reviewed by Samuelsen et al., 2006). The reason for this was for a long time unclear, however difficulties with identifying MHC class II and associated genes indicated changes in how Atlantic cod process classical MHC class II dependent antigens. Recently whole-genome sequencing revealed that the genome of Atlantic cod lack MHC class II and Invariant chain (Ii), and that CD4 is only present as a truncated pseudogene (Star et al., 2011). This would render the MHC class II antigen presenting pathway (Mantegazza et al., 2013) non-functional, and would explain the lack of specific antigen-responses when vaccinating with bacterins. Atlantic cod has expanded its repertoire of MHC class I antigens which might facilitate cross-presentation of traditional MHC class II antigens by MHC class I molecules, and there is evidence that Atlantic cod might be compensating for the loss of CD4+ T-cells as well by having different subsets of CD8+ T-cells (Star et al., 2011). Atlantic cod also has high levels of natural antibodies compared to other fish species (reviewed by Pilström et al., 2005), which might compensate for a strong specific antibody response on encounter with a pathogen. However, there are reports of Atlantic cod producing specific antibodies in response to vaccination with Aeromonas salmonicida, Listonella (Vibrio) anguillarum and F. noatunensis (Lund et al., 2006; Lund et al., 2007; Schrøder et al., 2009), though as they seem to be predominately recognizing LPS a T-cell independent antibody response (reviewed by Alugupalli, 2008) could explain the observed production of antibodies in response to these bacterial pathogens.
The production of membrane vesicles by cells is a conserved mechanism occurring throughout all domains of life, both prokaryotic and eukaryotic (reviewed by Deatherage & Cookson, 2012). In bacteria, these vesicles are usually called Outer Membrane Vesicles (OMVs) and are formed by budding from the outer bacterial membrane (from Gram negative bacteria). They are 10-300 nm in diameter and spherical, containing outer membrane and periplasmic proteins, and recent data indicates that they might contain inner membrane and cytoplasmic proteins as well, and in some cases DNA (Pèrez-Cruz et al., 2013+++). The protein content of OMVs show specific packaging, as some proteins are enriched and some are excluded (e.g. Galka et al., 2011; Haurat et al., 2011 og mange flere). The exact sorting mechanism responsible for enrichment or exclusion of proteins from OMVs is not currently known. Many pathogenic bacteria incorporate virulence factors, including toxins, into their OMVs, turning the vesicles into bacterial-derived bombs (Kuehn & Kesty, 2005; Galka et al., 2011; Haurat et al., and etec+salmonella as well). OMVs have recently received renewed focus in the field of vaccinology (reviewed by Collins, 2011), as they present antigens in their native conformation and does not require adjuvants to be immunogenic. Immunization of humans using OMVs have been performed with great success against Neisseria meningitidis type B (reviewed by Granoff, 2010; Collins, 2011). OMVs derived from other bacteria have also shown protective efficacy when used as vaccines against other pathogenic bacteria, such as Burkholderia pseudomallei (Nieves et al., 2011), Brucella melitensis (Avila-Calderon et al., 2012), Edwardsiella tarda (Park et al., 2011), enterotoxigenic Escherichia coli (Roy et al., 2011), Salmonella Typhimurium (Alaniz et al., 2007), Shigella flexneri (Camacho et al., 2011; Camacho et al., 2013) and Vibrio cholera (Schild et al., 2008). OMVs have been shown to induce both B- and T-cell responses (Alaniz et al., 2007. Romeu et al., 2013++)
F. tularensis ssp. has previously been shown to produce vesicles in in vitro cultured infected macrophages (Anthony et al., 1991; Golovliov et al., 2003). Recent work has shown that similar vesicles could be isolated from broth cultured F. tularensis ssp. novicida and F. philomiragia ssp. philomiragia (Pierson et al., 2011), and that these vesicles were derived by budding from the outer bacterial membrane (McCaig et al., 2013) thereby being true OMVs. Macrophages treated with the vesicles released proinflammatory cytokines, and mice vaccinated with OMVs were protected against subsequent challenge with F. tularensis ssp. novicida (Pierson et al., 2011; McCaig et al., 2013). Interestingly, in addition to regular spherical vesicles, OMVs from F. tularensis can also be shaped like tubes (McCaig et al., 2013). Previously, Bakkemo et al., (2011) showed by EM that also F. noatunensis ssp. noatunensis releases vesicles in vitro in infected macrophages, but as they could not detect vesicles from extracellular cultured bacteria, they hypothesized that the formation of vesicles from F. noatunensis ssp. noatunensis was an intracellular event.
Systems and methods for protecting fish against infection by infectious agents are needed.